Hi Stéphane,
On Tue, Dec 22, 2009 at 07:00:49PM +0100, Stéphane Chatty wrote:
> Hi Dmitry,
>
> I am working on the 3M driver and taking your comments into account (and
> applying them to the Stantum driver as well) but I have an question on
> the comment below:
>
>>
>>> +struct mmm_finger {
>>> + __s32 x, y;
>>> + __u8 rank;
>>> + int touch:1, valid:1;
>>> +};
>>
>> Does it make sense to turn access to touch and valid into
>> read-modify-write sequence? Just change them to be 'bool's, it won't
>> cause your structures to grow in size.
>>
>>> +struct mmm_data {
>>> + struct mmm_finger f[10];
>>> + __u8 curid, num;
>>> + int touch:1, valid:1;
>>
>> Same as above.
>
> Why do you think a read-modify-write sequence would be better? in my
> mind these were just cached values, just like x and y.
I meant "read-modify-write" is worse.
To modify a bitfiled the processor has to fetch data from memory (or
cache if it is in cache) perform a logical operation on it, and store
it back in teh memory. With boolean occupying 1 byte, on most
architectures (older alphas is one exception I know of) var=true can be
done by simple store, no need for "read-modify" part.
--
Dmitry
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
